1
Substrate’s effect on the establishment of rusty
crayfish (Orconectes rusticus)
Sam Pecoraro
BIOS 35502: Practicum in Field Biology
Advisor: Ashley Baldridge
2009
2
ABSTRACT:
The rusty crayfish is a well-studied invasive that has become a common
sight in Northern Wisconsin and the Upper Peninsula of Michigan. Likely a
result of human error, the negative effects of this species’ invasion can be
dramatic. Rusty crayfish have been found at very different densities in two lakes
associated with UNDERC property. One is a large, public access lake and the
other, only two kilometers away, is a small, very limited access lake. Given the
important influence substrate, especially cobble, can have on the population
dynamics and establishment of this species, I predicted that the substrates of the
two lakes would be different. Statistical testing of visual substrate
characterization from three depths revealed that the substrates of the two lakes
were indeed dissimilar. I also expected there to be a statistically higher numbers
of crayfish in the preferred, cobble habitat. Statistical tests revealed that there
were no significant associations between the number of crayfish captured from
trapping and hand collection with substrate as a main effect. Though my results
were not completely as expected, the differences in substrate between the two
lakes have important conservation implications for UNDERC.
INTRO:
Rusty crayfish (Orconectes rusticus) have become a common invasive
species on northern temperate lakes over the past 50 years. They may have been
3
introduced by the emptying of anglers’ bait buckets, commercial crayfish
harvesters, or property-owners trying to rid their lakes of weeds (Wilson et al.
2004). However benign the intentions of their introducers, the negative impacts
of O. rusticus on lake ecology have been long-studied and well documented (e.g.,
destruction of macrophytes, Lodge and Lorman 1987; Hanson and Chambers
1995; Nystrom and Strand 1996 ; Rosenthal et al. 2006; reduction of
macroinvertebrates, Wilson et al. 2004, McCarthy et al. 2006; displacement of
other crayfish, Olsen et al. 1991, Hill and Lodge 1994). Rusty crayfish are such a
proficient invasive because they typically out-compete and replace the native
species of crayfish O. propinquus and O. virilis (Hill and Lodge 1994). They are
then able to persist at abundances much higher (between 2-18 times as many
crayfish per trap) than their native counterparts for as long as ten years, which can
have whole food web implications (Wilson et al. 2004).
Substrate plays an important role in crayfish dynamics, especially species
establishment. According to Wilson et al. 2004, the availability of a cobble
habitat is an essential factor influencing O. rusticus establishment. Predation risk
for smaller crayfish tends to be lower in cobble habitats than in sand, muck or
macrophyte habitats because there is more debris for them to take refuge in
(Kershner and Lodge 1995). Additionally, cobble habitats are preferred by
female crayfish as nurseries (Lorman 1980). Rusty crayfish are able to displace
the other two species of crayfish from the preferential cobble habitat during the
4
day when predation risk is highest and cover is most important (Hill and Lodge
1994).
Orconectes rusticus have been discovered in two lakes on UNDERC
property, but seem to be establishing at much different rates. These two lakes
provide a unique opportunity to study the population dynamics of an important
invasive species. Data from the Lodge Lab indicate that while Orconectes
rusticus has been in Brown Lake for at least 5 years, it is only persisting at
relatively low population densities. One mature female was trapped at location 10
in Brown Lake in 2004, but none were captured over the next four years.
However, in July 2008, two male rusty crayfish were captured at trap location 10.
Both rusty crayfish had carapace lengths over 40mm, indicating that they are at
least Age 3. This is interesting, as it indicates that the Rusty crayfish population
is persisting, but below the requisite threshold for trapping.
In Big Lake, which is only about two km (as the crow flies) from Brown
Lake, Orconectes rusticus has existed since at least 1995. Rusty Crayfish have
since taken over in Big Lake, with trap catches up 24males/trap. No other species
of crayfish has been recorded in Big Lake since the Lodge Lab started logging
data in 2003.
Given the differences in populations of rusty crayfish between the lakes, I
hypothesize that there will be significant differences between the substrates in
Brown and Big Lakes, with Big containing a higher percentage of the preferred
5
cobble habitat. Additionally, since Rusty crayfish have reached the epidemic
proportions (>9 crayfish/trap) as described by Wilson et al. 2004, I believe Big
lake will have higher numbers of O. rusticus in all substrate types, with especially
high concentrations in the cobble.
METHODS:
I compared the substrates of both Brown and Big Lakes. There was a
good amount of data on the substrate of Big Lake from a previous
characterization by Brett and Jody Peters in 2007, so I mapped the littoral zone of
Brown Lake for comparison. I categorized the substrate as cobble, sand,
macrophyte, or muck. Visual substrate classification in Brown Lake was taken
every 50m along the shoreline from GPS data (41 points). Following the protocol
set by the Peters’ in 2007, I characterized substrate in Brown at three different
depths: 0.75m, ½secchi, and ¾secchi. Secchi depth on the Big Lake data was
measured to be 2.4m in 2007. My data on Brown Lake calculated secchi depth at
2.14m. A view box and/or an oar were used in deeper water to quantify the
substrate.
To assess current crayfish densities in both lakes for use in conjunction
with the substrate data, both hand collection and trapping techniques were used in
both Brown and Big lakes at three substrate types: cobble, sand, and macrophyte.
No trapping nor hand collection occurred in muck because crayfish so rarely
6
inhabit such areas. The areas were selected from the substrate characterization
with three plots of each substrate type for a sample size of nine. Hand collection
was performed in 0.25m2 quadrats set up randomly throughout areas of same
substrate type (four quadrats per substrate type per lake). Crayfish traps
(modified Gee minnow traps) were then set with 120g beef liver as bait in the
same substrate areas used for quadrat counting. Two traps were set in each of the
three substrate types per lake. Traps were needed because they tend to selectively
catch larger male crayfish, whereas hand counting more frequently catches
females. Species, carapace length, sex, and form were determined on captured
crayfish. Crayfish quadrat data was not used due to the large numbers without a
capture in Brown Lake and due to the relatively low numbers from Big Lake hand
collection data. Trapping data was averaged from the two trap catches at any
one plot. To continue the monitor trapping that helped make this research
possible, Brown Lake was trapped a second time at 20 trap locations. A KruskalWallis ANOVA test was performed using SYSTAT 12 software to test for
variation in crayfish catches with substrate as a main effect. I followed up with a
Friedman Analysis of Variance with sex and substrate as factors. Contingency
analysis was performed in R (version 2.8.1) to see if substrate distribution is
independent of lake.
RESULTS:
7
The substrates of Big and Brown Lakes were widely disparate. The
breakdown for Brown Lake was: 6% cobble, 71% vegetation, 14% sand, and 9%
muck. Big Lakes’ breakdown: 31% cobble, 12% vegetation, 36% sand, and 21%
muck (Fig. 1). The χ2 results showed that substrate distribution is not independent
of lake (χ2=73.61, df=3, p<0.001) (Fig. 5).
The trapping data was incontrovertibly different between the two lakes.
One O. virilis was captured during hand collection from Brown Lake and
seventeen Orconectes rusticus were collected on Big Lake’s quadrat searches.
One O. virilis was trapped on Brown Lake; 493 O. rusticus were trapped on Big
Lake (Fig. 2). Of the crayfish found on Brown Lake, one was a male found in a
cobble habitat and one was a female found in sand with carapace lengths of 25mm
and 43mm, respectively. Big Lake had 149 males and 361 females plus 5
individuals that escaped from hand collection before we had the chance to sex
them. The average carapace length for the rusty crayfish in Big Lake was
29.7mm ± 0.2mm SE.
The ANOVA that tested for substrate’s effect on number of crayfish
caught from the combined quadrat and trapping data was only performed on data
from Big Lake because of insufficient data for Brown (only two crayfish caught
total). Shapiro-Wilk and Levene tests for normality were run simultaneously
with the ANOVA, and revealed that the data violated the assumptions of the test.
The results of the subsequent Kruskal-Wallis test showed that there was no
8
statistically significant difference between the amount of crayfish caught in any
one substrate type (K-W=2.00, df=2, p=0.368; Fig. 3). One of the replicates from
the macrophyte substrate was a possible outlier with 71.5 average captures vs. 0.5
and 1.5 captures for the other two macrophyte replicates respectively. The next
highest average capture for any one replicate had a cobble substrate and had 39.5
crayfish. When the possible outlier was removed and the K-W Analysis of
Variance was run again, the results still had no statistical significance (K-W=4.25,
df=2, p=0.119).
No difference was found in trap capture by sex. The Friedman test for
difference in average trap capture on Big Lake by substrate and sex demonstrated
a statistically non-significant trend (Friedman stat.=3.0, df=1, p=0.083; Fig. 4).
DISCUSSION:
My predictions about substrate differences between the two lakes were
correct. Contingency analysis revealed that substrate distribution is actually
dependent on the lake. Big Lake had a much higher percentage (31%) of the
cobble habitat ideal to Rusty crayfish establishment than Brown Lake (6%) (Fig.
1). The macrophyte substrate type, however, was much higher in Brown (71%)
than in Big (12%) (Fig. 1). It is important to consider here that though
macrophyte habitats may offer some cover from predation for crayfish, they
somewhat paradoxically also harbor the very predators that crayfish hide from,
9
fish. The macrophyte communities of Big Lake could also be being altered by the
crayfish themselves. According to Wilson et al. 2004, rusty crayfish can
significantly reduce macrophyte densities. The sheer amount of vegetation
covering the littoral zone of Brown Lake, coupled with the large fish catches my
partner Grace Loppnow (Loppnow 2009) and I extracted from fyke and seine
netting, leads me to believe that top-down predator control could be a serious
factor in Orconectes rusticus’ inability to flourish in Brown Lake.
Contrary to my predictions, I found that average crayfish captures per site
did not differ by substrate in Big Lake (p=0.368; Fig. 3). These results were not
as expected. Even when the possible macrophyte outlier was dismissed our
results didn’t demonstrate even a marginally significant trend (p=0.119). The
very high trap catch at the outlying macrophyte site could have been skewed by
the substrate on either side of it. Both of the adjacent sites had been characterized
as cobble at two of the three classification depths. The close proximity of a
cobble substrate, which could serve as a crayfish source, could have influenced
the catch in this vegetated area. The small sample size could also be the cause of
the unexpected statistical insignificance of our results. A simple average by
substrate for Big Lake data showed that while cobble habitat still had the highest
average numbers of catches, sandy substrates had higher catches than macrophyte
areas (Fig. 3). A small sample size and variation among sites in each substrate
type are likely the cause of these counterintuitive results.
10
Though our statistics showed only a marginally significant trend
(p=0.083) for differences in sex by substrate, our data shows a large margin
between the number of males and females captured in traps on Big Lake. I think
we may have trapped at a unique time in the season for crayfish. The
overwhelming number of females captured in traps (361) compared to males
(149) indicates a possible molt transition in progress. I suspect that we trapped
right after gravid females had released their brood. The substantial energy
commitment requisite to brooding would have left them starving and the free meal
provided by the beef liver in our traps may have been irresistible. The relatively
small number of males found in our traps may be related to the fact that we only
found form two males in our traps. It is likely that the majority of the males were
in the process of molting when we did our trapping. It is hard to qualify this
without having trapped Big Lake previously or since we collected our data, but it
is nonetheless plausible that all of the males we trapped had just molted from
form one to form two. The remaining males were busy molting; this coupled with
the theory that the females were brooding explains the unexpectedly high number
of females captured in traps.
A high percentage of cobble substrate is paramount to O. rusticus
establishment according to the literature. Our data doesn’t conflict with that when
the trapping data is considered alongside the substrate data, though correlation
does not necessarily imply causation. The only significant result we obtained- the
11
dependence of substrate on lake, was predicted by my hypothesis. Big Lake’s
unambiguously higher number of crayfish captures fits within the framework of
disparate substrates. It is unfortunate that our data was not normally distributed,
and that parametric statistics could not be utilized.
It is important to note that, while we collected no O. rusticus during our
hand collection and trapping on Brown Lake, in continuing the Lodge Lab
monitor trapping that provided the context for our experimentation, we caught a
rusty crayfish at trap location 18. This is intriguing because it was on the
complete other side of the lake from the trap location 10, the only place where O.
rusticus has been found.
Rusty Crayfish have been found in Brown Lake over the course of past
five years, and we found an additional individual in a new location for the species.
Because of a low amount of cobble habitat and high levels of crayfish predators, I
am not convinced that the trajectory of the Rusty crayfish in Brown Lake will
parallel that of Big Lake. It is important to note that no species of crayfish has
been detected at high population densities in Brown Lake, though it’s true O.
rusticus can reach much higher populations than its congeners. The limited access
nature of Brown, which reduces the likelihood of additional introductions, and the
fact that O. rusticus has come no where near epidemic levels leads me to predict
that Brown Lake is not likely to sustain high population densities of rusty
crayfish.
12
ACKNOWLEDGEMENTS:
I would like to thank my partner, Grace Loppnow, whose help, support,
and camaraderie made this research possible. I would like to thank my mentor,
Ashley Baldridge, whose mentorship and hands-on help guided me throughout
my work. Thanks to Navit Reid for helping me characterize the substrate of
Brown Lake. Thanks to Heidi and Andy Mahon for their advice on my analysis
and results and help with quadrat searching. My hat’s off to Andrew Perry,
Derryl Miller, and Ashley Bozell for also helping me with quadrat searches. I am
grateful for the Lodge lab and its monitor trapping records, which gave context to
my work and Brett and Jody Peters who made substrate data on Big Lake readily
available to me. I would also like to thank the Hank family for making this
research program available to undergraduates like myself.
REFERENCES:
Hanson, J.M., and Chambers, P.A. 1995. Review of effects of variation in
crayfish abundance on macrophyte and macroinvertebrate communities of
lakes. ICES Mar. Sci. Symp. 199:175-182.
Hill, A. M., and D. M. Lodge. 1994. Diel changes in resource demand:
competition and predation in species replacement among crayfishes.
Ecology 75: 2118-2126.
13
Kershner, M.W., and Lodge, D.M. 1995. Effects of littoral habitat and fish
predation on the distribution of an exotic crayfish, Orconectes rusticus. J.
North Am. Benthol. Soc. 14: 414-422.
Lodge, D.M., and Lorman, J.G. 1987. Reductions in submersedmacrophyte
biomass and species richness by the crayfish Oronectes rusticus. Can. J.
Fish. Aquat. Sci. 44: 591-597.
Loppnow, G. L. 2009 The impact of fish predation on invasive Rusty Crayfish
(Orconectes rusticus) populations in two Northern temperate lakes.
UNDERC report, University of Notre Dame.
Lorman, J.G. 1980. Ecology of the crayfish Orconectes rusticus in northern
Wisconsin. Ph.D. thesis, University of Wisconsin-Madison.
McCarthy, J. M., C. L. Hein, J. D. Olden, and M. J. Vander Zanden. 2006.
Coupling long-tem studies with meta-analysis to investigate impacts of
non-native crayfish on zoobenthic communities. Freshwater Biology
51:224-235.
Nystrom, P., and Strand, J.A. 1996. Grazing by a native and and exotic crayfish
on aquatic macrophytes. Freshw. BioI. 36: 673682.
Olsen, T. M., D. M. Lodge, G. M. Capelli, and R. J. Houlihan. 1991.
Mechanisms of impact of an introduced crayfish (Orconectes rusticus) on
littoral congeners, snails, and macrophytes. Canadian Journal of Fisheries
and Aquatic Science 48:1853-1861.
Rosenthal, S. K., S. S. Stevens, and D. M. Lodge. 2006. Whole-lake effects of
invasive crayfish (Orconectes spp.) and the potential for restoration.
Canadian Journal of Fisheries and Aquatic Science 63:1276-1285.
Wilson, K. A., J. J. Magnuson, D. M. Lodge, A. M. Hill, T. K. Kratz, W. L. Perry,
and T. V. Willis. 2004. A long-term rusty crayfish (Orconectes rusticus)
invasion: Dispersal patterns and community change in a north temperate
lake. Canadian Journal of Fisheries and Aquatic Sciences 61: 2255-2266.
FIGURES:
14
Brown Lake substrate breakdown
Big Lake substrate breakdown
cobble
6%
veg
12%
muck
9%
cobble
31%
sand
14%
sand
36%
veg
71%
muck
21%
Figure 1.–breakdown of the four substrate types as characterized on Brown Lake
in 2009 and Big Lake in 2007. The cobble habitat essential to O. rusticus
establishment is much more prevalent in Big Lake.
Trapping Captures
493
500
450
400
350
Brown
300
Captures
Big
250
200
150
1
100
50
0
Lake
Figure 2.-Crayfish trapping data from Brown and Big Lakes. 18 total traps were
set in 9 different locations with three different substrate types in each lake.
15
Average captures by substrate type for Big Lake
60
50
Captures
40
Cobble
30
Sand
Veg
20
10
0
Substrate
Figure 3.-Distribution of captures on Big Lake by substrate type. Cobble has the
highest number of captures, as expected. Sand had more captures than vegetation,
however. No results statistically significant (p=0.368).
Sex Distribution by Substrate Type
100
90
80
Captures
70
60
M
50
F
40
30
20
10
0
Cobble
Sand
Veg
Substrate
Figure 4.-Distribution of capture by sex and substrate type. These are the sums
of the averages at each substrate type Cobble has the highest number of both
males and females, though no results statistically significant (p=0.083).
16
Figure 5.-Substrate map of Brown Lake created in Google© Earth from substrate
characterization at three different depths. The substrates of Brown and Big lakes
were statistically discrete (p<0.001).